Kling 4.0 Image to Video Guide: Animate a Photo

Kling 4.0 image to video turns a picture, like a product shot, portrait, illustration or room, into a moving clip. No image yet? Start with text to video.

How Kling 4.0 image to video works on kling4.run

There's no separate image mode. Add a picture beside the prompt box and the generator sends it as a visual reference:

• The image shows how things look: faces, outfits, products, art style, rooms.
• The prompt says what happens: what moves, how the camera travels, what must stay the same.

The image isn't locked in as the first frame, so the model may reframe the shot or re-render details. The output frame follows your aspect ratio setting, not the photo's shape. A prompt is still required: the generate button stays disabled until you write one.

For Kling AI's model specs, read What Is Kling 4.0?

Choose a source image that animates well

• One clear subject, sharp, large in the frame and not cut off.
• Even light. Soft window or studio light beats hard flash and heavy filters.
• The best original, not a screenshot or compressed chat copy. The 300 px minimum is a floor, not a target.
• A calm background. Clutter invites motion you didn't ask for.
• Your output frame. Crop to the aspect ratio you'll generate; the default is 16:9.
• One scene. Collages and before-and-after splits confuse the model.

Supported image formats and limits

Files are checked when you pick them, and the uploader explains any fix; nothing uploads until you press generate. Chrome and some other browsers can't read TIFF files, so if a TIFF is rejected, save it as PNG or JPG. All six output ratios fit the shape limit, so a photo cropped to your frame passes that check.

Turn an image into a video, step by step

1. Open the AI video generator. Generating requires a free account; until you sign in, the button reads Sign in.
2. Click Add reference, left of the prompt, and choose your image.
3. Write the motion prompt. Type @ and pick Image 1 to point at it.
4. Open the settings button under the prompt to set aspect ratio, video length and resolution.
5. Check the cost (✦ plus a number) beside the generate button, then press it. Credits are deducted when the task starts; top up on the pricing page.
6. Once "Generation started" appears, you can leave the page. Most videos finish in two to five minutes and are saved to My Videos for MP4 download.

Longer, higher-resolution videos cost more, so test motion short and at low resolution first; re-runs won't match exactly.

Write motion prompts for a still image

The image covers appearance, so spend your words on change:

1. Subject motion: one action with direction and speed ("turns slowly toward the camera," not "moves").
2. Camera move: one move, or "static camera."
3. What stays the same: details that must survive, like a logo or hairstyle.

There's no negative-prompt field, so describe the result you want instead of listing what to avoid.

Product shot

Image: a speckled mug on a café table in window light.

Steam curls up from the coffee as sunlight slides across the table. Slow push-in. Keep the glaze and small logo from @image1 unchanged.

Portrait

Image: a gray-bearded man in a green scarf, looking aside, snow falling.

He turns toward the camera, blinks and smiles slightly as snowflakes drift past. Slow push-in to a close-up. Keep his face, beard and scarf from @image1.

Illustration

Image: a flat vector hot-air balloon over patchwork fields.

The balloon rises and drifts right while clouds slide past in layers. Static camera. Keep the flat vector style, palette and clean outlines from @image1.

Interior

Image: an empty loft kitchen with a tall window and oak cabinets.

Slow dolly toward the window as curtains sway and sunlight shifts across the counter. Keep the layout and cabinet colors from @image1. The kitchen stays empty throughout.

Combine several images with @mentions

Use several images when one photo can't carry everything, such as a character and a location. Images get lowercase tokens in upload order (@image1, @image2), shown in the prompt as badges labeled "Image 1" and "Image 2." Give each one a role:

The woman from @image1 walks through the greenhouse in @image2 and looks up at a tall palm. Tracking shot. Keep her face and jacket from @image1 and the glass roof from @image2.

• Hover over the reference stack to see each label.
• Removing an image renumbers the ones after it, but mentions in your prompt don't update. Re-insert them.

For video clips or audio, see reference to video.

What image to video is good for

• Product shots: use a photo with the label facing the camera, then ask for an orbit, a light sweep or rising steam.
• Portraits and characters: pick a photo with the eyes visible, then ask for a head turn, a blink or hair moving in the wind.
• Artwork and illustrations: upload the full-resolution file, then ask for parallax, moving water or drifting clouds.
• Real estate and travel: start from a level, well-lit wide shot, then ask for a slow dolly or a gentle pan.

Troubleshooting image to video results

If the content filter blocks a request, edit the prompt or try a different image. Still stuck? Email support@kling4.run.

Related guides

• Reference to video: images plus video clips and audio.
• Prompt guide: camera language, lighting and pacing.
• Credits and troubleshooting: costs and failed runs.
• Ready? Open the AI video generator.